2.1 Characterizations
The SEM images of MIL-68(Fe), g-C3N4, and MIL-68(Fe)/g-C3N4 are shown in Fig. 2. The theoretical framework configuration of MIL-68(Fe) is a three-dimensional network, consisting of one-dimensional channels of triangular and hexagonal types along the c-axis. The SEM images reveal that the prepared MIL-68(Fe) is composed of large, coarse polyhedral prisms, consistent with previous reports(Liang et al. 2021). However, its average size reaches the micron scale (Fig. 2-1a), and such large dimensions could likely impact the catalytic performance of MIL-68(Fe).
In the SEM images of the MIL-68(Fe)/g-C3N4 composite, g-C3N4 and MIL-68(Fe) are observed to be closely integrated. Additionally, the size of MIL-68(Fe) in the MIL-68(Fe)/g-C3N4 composite is much smaller than that of the original MIL-68(Fe), reducing from the micron scale to the nanometer scale. This indicates that g-C3N4, as a precursor, plays a crucial role in regulating the formation of MIL-68(Fe) crystals during synthesis. Although the size of the composite material is reduced, all samples retain the three-dimensional network structure of MIL-68(Fe). The size reduction of MIL-68(Fe)/g-C3N4-3 is the most significant, showing a needle-like structure with a width of only a few tens of nanometers under the microscope. The reduction in size brought by the composite is beneficial for increasing the specific surface area of the material, thereby providing more exposed surfaces and active sites.
The elemental distribution on the surface of the materials was analyzed using EDS. Figure 3 shows the EDS elemental mapping of the representative composite material MIL-68(Fe)/g-C3N4-3. The elemental maps reveal the presence of four elements: C, N, O, and Fe. The brighter areas corresponding to Fe and O indicate higher concentrations, originating from MIL-68(Fe) on the surface of g-C3N4, while the weaker N signal comes from g-C3N4. These results demonstrate that MIL-68(Fe) is highly dispersed on the surface of the composite material, which is beneficial for exposing more active sites in the MIL-68(Fe)/g-C3N4-3 composite.
The XRD patterns of MIL-68(Fe), g-C3N4, and the prepared MIL-68(Fe)/g-C3N4 composites are shown in Fig. 4. The characteristic diffraction peaks of the MIL-68(Fe)/g-C3N4 composites align well with those reported for MIL-68(Fe)(Liang et al. 2017)and g-C3N4(H. Li et al. 2014), The intensity of the MIL-68(Fe) characteristic peaks in the composites is comparable to that of the original MIL-68(Fe), indicating that the crystal structure of MIL-68(Fe) is preserved in the composites. This is logical due to two reasons: (1) the percentage of g-C3N4 in the MIL-68(Fe)/g-C3N4 composites is relatively small, and (2) the in situ growth process ensures a uniform distribution of MIL-68(Fe) on g-C3N4, explaining the absence of g-C3N4 diffraction peaks in composites with low g-C3N4 content, such as MIL-68(Fe)/g-C3N4-3, MIL-68(Fe)/g-C3N4-4, and MIL-68(Fe)/g-C3N4-5. Additionally, a shift in the characteristic peaks of g-C3N4 in the composites suggests strong coupling interactions between g-C3N4 and MIL-68(Fe) during the formation of the heterojunction.
Figure 5 shows the FTIR spectra of MIL-68(Fe), g-C3N4, and MIL-68(Fe)/g-C3N4-3. The prepared MIL-68(Fe)/g-C3N4-3 exhibits similar characteristic peaks to MIL-68(Fe). The peak at 555 cm⁻¹ corresponds to the Fe-O stretching vibration, indicating the formation of metal-oxygen bonds between the carboxyl group of terephthalic acid and Fe(Ⅲ)(C.-C. Wang et al. 2016). The peak at 746 cm⁻¹ corresponds to the C-H bond vibration in the benzene ring. The absorption bands of carboxyl groups coordinated with the metal center (C = O, C-O) are located at 1545 cm⁻¹ and 1392 cm⁻¹(Giannakoudakis et al. 2017), respectively. Additionally, the characteristic peaks at 1236 cm⁻¹ and 1641 cm⁻¹ might overlap with some peaks of g-C3N4, causing a shift in the characteristic peaks of g-C3N4 in the composite. Therefore, the FTIR results further confirm the formation of strong interaction heterojunctions in the MIL-68(Fe)/g-C3N4-3 material.
Furthermore, the chemical composition of g-C3N4 nanosheets and original g-C3N4 was analyzed using XPS. According to the XPS spectra in Fig. 6, the MIL-101(Fe)/g-C3N4 heterostructure contains four elements: C, N, Fe, and O. The N 1s characteristic peaks in the range of 396–402 eV (Fig. 6b) further confirm the successful incorporation of g-C3N4 into the heterostructure. The two peaks at 400.8 and 398.6 eV in the N 1s spectrum correspond to sp² hybridized N atoms in C-N and C = N, respectively. In the C 1s spectrum (Fig. 6b), the peaks at 284.0 and 285.4 eV correspond to C-C and C-N groups in g-C3N4, and the peaks at 284.7 and 288.5 eV correspond to the benzoic acid and C = O bonds in H2BDC. In the O 1s spectrum (Fig. 6d), the peaks at 531.8 eV and 530.2 eV correspond to the oxygen components of terephthalic acid and the Fe-O bonds in MIL-68(Fe). The Fe 2p XPS spectrum (Fig. 6e) shows two peaks at 711.9 and 725.6 eV, corresponding to Fe 2p₃/₂ and Fe 2p₁/₂, respectively. The BE difference of 13.7 eV between these peaks indicates that Fe is in the + 3 oxidation state, which is further supported by the peak at 716.4 eV.
The thermal stability of MIL-68(Fe)/g-C3N4-3, MIL-68(Fe), and g-C3N4 in a nitrogen atmosphere was investigated using thermogravimetric analysis (TGA). As shown in the TGA curves in Fig. 7b, the original MIL-68(Fe) exhibited only minimal weight loss in the temperature range of 25–300°C, but significant weight loss occurred at temperatures above 300°C, indicating that MIL-68(Fe) does not possess particularly good thermal stability. Pure g-C3N4 showed very slight weight loss throughout the entire temperature range, with noticeable weight loss starting above 600°C, demonstrating good thermal stability. Given the proportion of g-C3N4 added in the preparation method, g-C3N4 constitutes only about 1% of MIL-68(Fe)/g-C3N4-3, so the weight retention of pure g-C3N4 has a negligible impact on the weight change of MIL-68(Fe)/g-C3N4-3 upon heating. MIL-68(Fe)/g-C3N4-3 experienced minimal weight loss in the temperature range of 25–420°C, with significant weight loss starting above 420°C. After the composite material was formed, MIL-68(Fe)/g-C3N4-3 exhibited better thermal stability than the original MIL-68(Fe).
The UV-vis characterization of MIL-68(Fe)/g-C3N4-3 and MIL-68(Fe) (Fig. 7a) shows that MIL-68(Fe)/g-C3N4-3 has a UV-vis spectrum that highly overlaps with that of MIL-68(Fe), indicating that the optical properties of g-C3N4 nanosheets did not change. This preservation of the optical properties is beneficial for maintaining the photocatalytic performance of MIL-68(Fe)/g-C3N4-3.
2.2 Photocatalytic performance
The prepared MIL-68(Fe)/g-C3N4 material was used as a photocatalyst for pollutant removal, and its photocatalytic performance was investigated. First, the photocatalytic reduction of Cr(Ⅵ) by MIL-68(Fe)/g-C3N4 was studied. To eliminate potential effects from hydrolysis or photolysis, control experiments were conducted. As shown in Fig. 8, the concentration of Cr(Ⅵ) did not decrease in the absence of a catalyst, indicating that Cr(Ⅵ) alone does not react under sunlight irradiation without a photocatalyst.
Before conducting the photocatalytic reaction with the catalyst, the adsorption of Cr(Ⅵ) by the photocatalyst in the dark was explored. Under dark conditions, g-C3N4, MIL-68(Fe), and MIL-68(Fe)/g-C3N4 could all adsorb small amounts of Cr(Ⅵ) and reached equilibrium within 40 minutes. According to previous research, MIL-68 achieved a Cr(Ⅵ) removal rate of 76.0% when ammonium oxalate was used as a scavenger, but only an 8% reduction rate for Cr(Ⅵ) without the scavenger(Jing et al. 2017). In our experiments, which did not use a scavenger, MIL-68(Fe) achieved a similarly low Cr(Ⅵ) reduction rate of 8%. g-C3N4 also exhibited a low Cr(Ⅵ) reduction rate, consistent with previous reports.
In contrast, the MIL-68(Fe)/g-C3N4 heterojunction photocatalyst demonstrated higher photocatalytic activity for Cr(Ⅵ) reduction compared to the individual materials. Among all the photocatalysts, MIL-68(Fe)/g-C3N4-3 showed the highest Cr(Ⅵ) removal efficiency, reducing the Cr(Ⅵ) concentration by 99.9% after 80 minutes of reaction. This high efficiency is attributed to the small needle-like morphology and large specific surface area of MIL-68(Fe)/g-C3N4-3.
The photocatalytic reduction kinetics of Cr(Ⅵ) are shown in Fig. 9. and the samples followed a first-order kinetic model. The study of photocatalytic Cr(Ⅵ) reduction kinetics revealed that the reduction rate of Cr(Ⅵ) by MIL-68(Fe)/g-C3N4-3 was 0.04519 min⁻¹, which is approximately 376 times higher than that of MIL-68(Fe) alone (0.00012 min⁻¹) and 594 times higher than that of g-C3N4 alone (0.000076 min⁻¹). Additionally, the physical mixture of MIL-68(Fe) and g-C3N4 exhibited poor Cr(Ⅵ) reduction efficiency, significantly lower than that of MIL-68(Fe)/g-C3N4-3. These results indicate that the introduction of g-C3N4 into MIL-68(Fe) significantly enhances the photocatalytic activity of MIL-68(Fe)/g-C3N4-3.
The photocatalytic degradation of tetracycline by MIL-68(Fe)/g-C3N4 was also tested. To eliminate the possibility of natural photolysis, control experiments were conducted. As shown in Fig. 10, the concentration of tetracycline did not decrease under light irradiation without a photocatalyst, indicating that light alone does not degrade tetracycline. Before starting the photocatalytic degradation reaction, the adsorption of tetracycline by the photocatalysts in the dark was investigated. Under dark conditions, all the catalysts adsorbed tetracycline within 40 minutes, with MIL-68(Fe)/g-C3N4-3 showing the highest adsorption amount due to its small needle-like morphology and large specific surface area.
When using the original MIL-68(Fe) as a photocatalyst, the degradation rate of tetracycline was only 29.6%, indicating weak photocatalytic degradation ability. The g-C3N4 photocatalyst exhibited a high photocatalytic activity for tetracycline degradation, achieving a degradation efficiency of 92% after 40 minutes of light irradiation. In comparison, the degradation rate of tetracycline by MIL-68(Fe)/g-C3N4 was significantly higher than that of MIL-68(Fe) alone but did not exceed that of g-C3N4 alone.
According to the synthesis method of the composite materials, the amount of g-C3N4 added was very small, approximately 1% in MIL-68(Fe)/g-C3N4-3. Therefore, the catalytic ability of g-C3N4 alone has a negligible impact on the catalytic performance of MIL-68(Fe)/g-C3N4-3. The enhanced performance of MIL-68(Fe)/g-C3N4 in photocatalytic degradation of tetracycline is not due to the excellent degradation ability of g-C3N4 itself but rather because the incorporation of g-C3N4 into MIL-68(Fe) forms a scaffold structure that improves the morphology and specific surface area, thereby accelerating the photocatalytic degradation of tetracycline.
Among the MIL-68(Fe)/g-C3N4 composites, MIL-68(Fe)/g-C3N4-3 showed the highest removal efficiency for tetracycline, achieving a degradation rate of 90% after 80 minutes of light irradiation. The optimal photocatalytic degradation result of tetracycline by MIL-68(Fe)/g-C3N4-3 is consistent with the results of photocatalytic reduction of Cr(Ⅵ). MIL-68(Fe)/g-C3N4-3 can be considered the optimal composite material for photocatalytic removal of pollutants.
In actual industrial environments, wastewater often contains heavy metals along with other harmful organic pollutants. The simultaneous oxidation and reduction reactions of photocatalysts to remove pollutants are of significant importance in industrial applications. As shown in Fig. 12a, in the absence of a photocatalyst, the concentrations of tetracycline and Cr(Ⅵ) remained essentially unchanged after 160 minutes of light irradiation when both were present. However, when MIL-68(Fe)/g-C3N4-3 was used as a photocatalyst, the reduction efficiency of Cr(Ⅵ) greatly increased in the presence of tetracycline, with Cr(Ⅵ) almost completely reduced after 40 minutes of light irradiation. In the absence of tetracycline, the reduction rate of Cr(Ⅵ) was only 48%
As shown in Fig. 12b, the presence of Cr(Ⅵ) also significantly improved the degradation efficiency of tetracycline by MIL-68(Fe)/g-C3N4-3. After 40 minutes of reaction, the removal rate of tetracycline reached 70%, compared to only 53% in the absence of Cr(Ⅵ) under the same reaction time. However, as the reaction proceeded and Cr(Ⅵ) was nearly completely reduced, the reaction rate in the group with added Cr(Ⅵ) slowed down, and eventually, both groups achieved approximately the same removal rate of tetracycline.
These results highlight the enhanced photocatalytic performance of MIL-68(Fe)/g-C3N4-3 in the simultaneous removal of both Cr(Ⅵ) and tetracycline, demonstrating its potential effectiveness in treating industrial wastewater containing multiple types of contaminants.
The stability of photocatalysts is crucial for practical applications. Therefore, MIL-68(Fe)/g-C3N4-3 was subjected to five consecutive cycles of photocatalytic removal of Cr(VI) in a coexistence environment to assess its stability. As shown in Fig. 13, MIL-68(Fe)/g-C3N4-3 maintained good photocatalytic activity after five cycles. Interestingly, the reduction efficiency of Cr(VI) by MIL-68(Fe)/g-C3N4-3 did not change significantly during the first three cycles but slightly decreased in the last two cycles. This could be due to Cr(III) adsorbing onto the surface of MIL-68(Fe)/g-C3N4-3 during the previous reaction, where the disproportionation reaction between adsorbed Cr(III) and Cr(VI) accelerated the reduction of Cr(VI). However, as the number of cycles increased, the excessive adsorption of Cr(III) on the surface of MIL-68(Fe)/g-C3N4-3 covered the active sites.
The chemical stability of the recovered MIL-68(Fe)/g-C3N4-3 was also tested using XRD. In Fig. 14a. the XRD patterns of MIL-68(Fe)/g-C3N4-3 showed little change after five cycles, indicating that MIL-68(Fe)/g-C3N4-3 has good chemical stability. In Fig. 14b SEM images further confirmed that MIL-68(Fe)/g-C3N4-3 maintained its structure well with no significant changes after five cycles.
2.3 Exploration of the possible Mechanisms for Enhanced Photocatalytic Performance
As shown in Fig. 16, the N2 adsorption/desorption isotherms and pore size distribution of the samples were analyzed. According to previous studies (Tan et al. 2017), pure g-C3N4, being a non-porous material, exhibits relatively low specific surface area and porosity. The original MIL-68(Fe) also showed a relatively low specific surface area and pore size (SBET 186.9 m²/g). Compared to other Fe-MOFs, MIL-68(Fe) has a lower specific surface area due to its bulky structure, limiting its catalytic applications.
After forming the composite, MIL-68(Fe)/g-C3N4-3 exhibited a nearly five-fold increase in specific surface area compared to MIL-68(Fe), consistent with the reduced size of the composite observed in SEM characterization. g-C3N4 provides a template for the growth of MIL-68(Fe), which cuts the bulky structure of MIL-68(Fe), enhancing the exposure of the porous layers within the crystal. Some Fe-MOF/g-C3N4 composites have shown decreased specific surface areas after formation, such as MIL-100(Fe)/g-C3N4 (from 198.7 m²/g to 131.6 m²/g) and MIL-53(Fe)/g-C3N4 (from 20.6 m²/g to 18.5 m²/g). However, the synthesized MIL-68(Fe)/g-C3N4-3 displayed an impressive increase in specific surface area (from 186.9 m²/g to 1046.8 m²/g). This significant increase in specific surface area is attributed to the g-C3N4's cutting effect, which prevents the bulk accumulation of MIL-68(Fe) and promotes the growth of smaller crystals without destroying the basic structure of MIL-68(Fe).
The light absorption properties of semiconductors determine their photocatalytic activity. Therefore, UV-DRS analysis of MIL-68(Fe) and MIL-68(Fe)/g-C3N4-3 was conducted, as shown in Fig. 15. The original MIL-68(Fe) can absorb visible light, appearing orange. The calculated band gap of MIL-68(Fe) from the Kubelka-Munk plot is 2.18 eV. After doping with g-C3N4, the UV absorption of MIL-68(Fe)/g-C3N4-3 decreased. The absorption edge of the composite showed a red shift, indicating that the band gap of the MIL-68(Fe)/g-C3N4 heterostructure narrowed. The band gap of MIL-68(Fe)/g-C3N4-3, calculated from the Kubelka-Munk plot (Fig. 15b), is 1.98 eV, suggesting that MIL-68(Fe)/g-C3N4-3 may have better visible light response. Similar band gap reductions have been observed in other MOFs heterojunctions with rGO(Yang et al. 2017)和g-C3N4(Huang et al. 2017), indicating that narrower band gaps facilitate improved photocatalytic efficiency.